. Weihl, W.E.: Commutativity-based concurrency control for abstract data types. IEEE Transactions on Computers, Vol.37, No.12, pages 1488-1505, December 1988  Home/Search   Document Not in Database   Summary   Related Articles   Check

....work in work ows with a focus on correctness issues. 2.1 Non serializability correctness criterion Serializability has been widely accepted as the correctness criterion for transaction processing systems. Although the semantics of abstract operations can be used to narrow the notion of a con ict [71], 34] 6] 72] serializability is still the goal of these approaches. To enhance performance and to accommodate some advanced applications in which long running activities might be present, non serializability correctness criterion are proposed in [28] 49] 27] and [46] In these models, ....

W. Weihl. Commutativity-based concurrency control for abstract data types. IEEE Trans. on Computers, 37(12), Dec. 1988.

....itself. Second, the trec is never left in an incorrect state, so that search suboperations do not need to be blocked. Third, many of the modifying suboperations commute. Several authors have studied the issue of concurrency control on abstract data types when some operations may commute [31, 32, 33, 34, 35], but in the context of a transaction processing system. Our algorithms are similar to those described by Ellis [25] to maintain the replicated directories of the distributed hash tables in so that it is not always necessary that all 2.2 Integrating Concurrency Control with Replica Coherency ....

W.E. Weihl. Commutativity-based concurrency control for abstract data types. IEEE Transactions on Computers, 37(12):1488-1505, 1988.

....based on their semantics. This level of reasoning is characteristic of that needed to prove the correctness of a schedule built on consistency assertions. Commutativity tables were introduced in the late 80s as a way to specify semantics information for enhanced concurrency control algorithms [Wei88] Commutativity tables specify boolean conditions under which two atomic database operations commute (do not conflict) Di#erent kinds 11 of commutativity (forward, backward, right backward) have been identified and serve di#erent purposes in advanced transaction models. Forward and backward ....

William E. Weihl. Commutativity-based concurrency control for abstract data types. IEEE Transactions on Computers, 37(12):1488--1505, December 1988.

....kind of firm serialization causes dependencies among requests. If some users abort already executed requests, all subsequently executed requests must be undone and redone to invalidate the effects of those aborted requests. These dependencies can be weakened by commutative relations among methods[7]. However, in general, it is difficult to define commutativity relations among methods, and it is necessary to evaluate the different orders in which methods can be executed to increase the commutativity relations [6] One solution to these problems involves the use of fine grained objects, and on ....

William E. Weihl. Commutativity-Based Concurrency Control for Abstract Data Types. IEEE Transaction on Computer, 37(12):1488--1505, Dec. 1988.

....of Variable x 2 Temporal Operator Always 3 Temporal Operator Eventually flflfl Temporal Operator Next op Operation Table 1 Relevant Object Z Notation 4 Chapter 1 2 RELATED WORK The work on semantic based concurrency control can be classified into two major categories. In the first category [10, 11, 15, 16] the authors exploit the semantics of operations to increase concurrency. Instead of using low level database operations like read or write to access the database objects, the authors propose using high level operations to access these objects. Commutativity of these operations, and not the ....

W. E. Weihl. Commutativity-based concurrency control for abstract data types. IEEE Transactions on Computers, 37(12):1488--1505, December 1988.

....the need arises. 2 Related Work Most transaction oriented models enforce a very low level, syntactic notion of consistency, namely serializability with respect to read write conflicts [3] Two separate lines of work have expanded on this notion. The first line is the work on atomic transactions [8, 9, 13, 14, 20, 21]. This line of work is based on ADTs. We also adopt the theory of ADTs to define correctness, but there is a crucial difference in focus. We are interested in achieving more concurrency by expanding the set of correct execution histories such that some transactions need not be atomic. The second ....

William E. Weihl. Commutativity-based concurrency control for abstract data types. IEEE Transactions on Computers, 37(12):1488--1505, December 1988. 11

....is composed of a set of design patterns that describes its architecture and, at the same time, documents how to use TransLib. The two main design patterns of TransLib are TransLock for concurrency control and TransRecovery for recovery. TransLib provides commutativity based concurrency control [3] which allows more concurrency than the traditional read write locking, as well as several predefined recovery mechanisms. One of the contributions of TransLib is to allow the redefinition of recovery and concurrency control mechanisms. This feature allows customizing the framework to the ....

....based locking and recoverability. Read write locking was the first proposed scheme, and it is the less concurrent of the three. It provides exclusive writes and concurrent reads. The other two schemes are based on user defined locks to provide more concurrency. In commutativitybased locking [3] two operations conflict only if they do not commute. Read write locking is a special case of commutativity based locking. Recoverability [14] provides yet even more concurrency, as two operations only conflict, if the second depends on the result of the first. However, recoverability increases ....

B. Weihl. Commutativity Based Concurrency Control for Abstract Data Types. IEEE Transactions on Computers, 37(12):1488--1505, December 1988.

....programs, it may also benefit other areas of computer science. For example, commuting operations allow computations on persistent data in object oriented databases to execute in parallel. Transaction processing systems can exploit commuting operations to use more efficient locking algorithms [Weihl 1988]. Commuting operations make protocols from distributed systems easier to implement efficiently; the corresponding reduction in the size of the associated state space may make it easier to verify the correctness of the protocol. In all of these cases, the system relies on commuting operations for ....

WEIHL, W. 1988. Commutativity-based concurrency control for abstract data types. IEEE Transactions on Computers 37, 12 (Dec.), 1488--1505.

....approaches are to impose restrictions on global transactions or to release the requirement of serializability. 7] The restrictions include confining the operations of global transactions to read and restricting the way of submitting global transactions to be serial. Other researches, such as [8, 5, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18] also lead their way to releasing the global serializability. In the next section, we propose a new correctness criterion other than global serializability (the criterion used in the past researches) to resolve the problem caused by indirect conflicts. 3 Global Concurrency Controls Instead of ....

W. E. Weihl, Commutativity-Based Concurrency Control for Abstract Data Types, IEEE Transactions on Computers, Vol. 37, No. 12, December 1988.

....the representation of an object can be changed without affecting the rest of the database, and users do not have to know the structure, methods and constraints applicable to objects. Third, OO DBMSs present an opportunity to provide more concurrency than traditional approaches allow [Weih88]. In the object oriented approach, the database system knows more about the operations being performed. They are not simply read or write operations, but rather have more semantics. The concurrency of transactions executing on objects can be enhanced through the use of the semantic information ....

.... interleaved to preserve data consistency; some of these techniques are presented in [Garc83, Lync83] Object based semantic concurrency control techniques allow the users to specify allowable interleaving of object operations to preserve logical consistency of data; some of these are presented in [Badr88, Badr92, Schw84, Weih88, Wolf93]. The studies in [Wolf92, Wolf93b, Wolf93c] focused on semantic based concurrency control mechanism for an object oriented real time database system. They presented a method for dynamically determining compatibility relations of operations by using a run time compatibility function for each pair ....

Weihl, W., "Commutativity-Based Concurrency Control for Abstract Data Types," IEEE Transactions on Computers, 37(12):1488-1505, December 1988.

....Protocols Natural Easy to Use API Weight Spec. Conit based Continuous Consistency Model Conit Theory Practicality Generality Figure 1. How the conit based consistency model achieves two typically conflicting goals. In the context of traditional replicated databases, much research [2, 3, 4, 8, 9, 10, 18, 19, 20, 24, 25, 26, 30, 31, 32] has been performed on relaxed consistency models. However, such traditional models typically achieve only one of generality and practicality. Some of the consistency models [2, 19, 20, 31] are general enough to allow a wide range of applications to express their consistency semantics. However, ....

....range of applications to express their consistency semantics. However, they provide no practical, efficient, applicationindependent protocols to enforce the model and no natural API for application programmers, thus failing to meet the practicality requirement. Other relaxed consistency models [3, 4, 8, 9, 10, 18, 24, 25, 26, 30, 32] have easy to use interfaces and can be efficiently implemented, but they typically address the consistency requirements of a specific class of applications. In this paper, we propose a conit based continuous consistency model for wide area data replication to simultaneously achieve generality ....

[Article contains additional citation context not shown here]

W. E. Weihl. Commutativity-Based Concurrency Control for Abstract Data Types. IEEE Transactions on Computers, December 1988.

....Protocols Natural Easy to Use API Weight Spec. Conit based Continuous Consistency Model Conit Theory Practicality Generality Figure 1. How the conit based consistency model achieves two typically conflicting goals. In the context of traditional replicated databases, much research [2, 3, 4, 7, 8, 9, 15, 16, 17, 20, 21, 22, 25, 26] has been performed on relaxed consistency models. However, such traditional models typically achieve only one of generality and practicality. Some of the consistency models [2, 16, 17, 26] are general enough to allow a wide range of applications to express their consistency semantics. However, ....

....range of applications to express their consistency semantics. However, they provide no practical, efficient, applicationindependent protocols to enforce the model and no natural API for application programmers, thus failing to meet the practicality requirement. Other relaxed consistency models [3, 4, 7, 8, 9, 15, 20, 21, 22, 25] have easy to use interfaces and can be efficiently implemented, but they typically address the consistency requirements of a specific class of applications. In this paper, we propose a conit based continuous consistency model for wide area data replication to simultaneously achieve generality ....

[Article contains additional citation context not shown here]

W. E. Weihl. Commutativity-Based Concurrency Control for Abstract Data Types. IEEE Transactions on Computers, December 1988.

....as necessary. Table 1 briefly explains the Z notation used in our examples. 2 Related Work Most transaction oriented models enforce a very low level, syntactic notion of consistency, namely serializability with respect to read write conflicts [BHG87] An expansion is the atomic transactions work [Her87, HW91, LMWF94, Lyn83, Wei84, Wei88], in which access operations are given by the particular abstract data type. We relax the requirement that transactions in correct executions histories appear atomic. Many researchers have broken transactions into steps and developed semantics based correctness criteria for decompositions ....

W.E. Weihl. Commutativity-based concurrency control for abstract data types. IEEE Transactions on Computers, 37(12):1488--1505, December 1988.

....The semantics based mechanisms can be divided into two main groups [Skarra and Zdonik 32 89] One group defines concurrency properties on abstract data types, while the other defines concurrency properties on the transactions themselves. The first group of mechanisms [Weihl 88; Herlihy and Weihl 88] constrains interleaving of concurrent transactions by considering conflicts between operations defined on typed objects. Describing the details of this group of mechanisms requires an overview of abstract data types and object oriented systems. But since these concepts are outside the scope of ....

Weihl, W. Commutativity-Based Concurrency Control for Abstract Data Types (Preliminary Report). Proceedings of the 21st Annual Hawaii International Conference on System Sciences, IEEE Computer Society Press, January, 1988, pp. 205-214.

....and write. For example, object oriented databases use abstract data type techniques to define data objects which support specific and rather complex operations (see, e.g. ZM90] Having semanticallyricher operations allows refining the notion of conflicting versus commutative operations [BR87, Wei88] That is, it is possible to examine whether two operations commute (i.e. do not conflict) and hence can be executed concurrently. By contrast, in the classical model, there is not much scope for such considerations since a write operation conflicts with any other operation on the same entity. ....

....Y , commute, if (X ffi Y ) j (Y ffi X) Two operations conflict if they do not commute. Observe that defining operations as functions, regardless to whether they read or update the database, leads to a very simple definition of the key concept of commutativity. Compare our definition to those of [Wei88, BR87] for example) Typically, an operation that updates an entity and an operation that reads it do not commute. Part of the orderings implied by the total order in which operations are composed to form a history are arbitrary, since only conflicting operations must be totally ordered. In ....

[Article contains additional citation context not shown here]

W. E. Weihl. Commutativity-based concurrency control for abstract data types. IEEE Transactions on Computers, C-37(12):1488--1505, December 1988.

.... transaction and data logical consistency; some of these techniques are presented in [GM83, Lyn83] Object based semantic concurrency control techniques allow the user to specify allowable interleavings of object operations to preserve data logical consistency; some of these are presented in [BR88, BR92, Wei88, WDL93, SS84]. However, these techniques typically only support limited forms of logical consistency; they do not address temporal consistency or the tradeoffs among the four RTCC requirements. In Section 2 we present a general model of an object oriented real time database system that we use to describe our ....

William Weihl. Commutativity-based concurrency control for abstract data types. IEEE Transactions on Computers, 37(12):1488--1505, December 1988.

....control in transaction systems today. In recent years much research has focused on extending concurrency control methods to take the semantics of the data into account, thus permitting more concurrency by allowing transactions executing commuting operations to run concurrently (e.g. see [8, 13, 16, 15, 14]) Such logical locking can be important to avoid concurrency bottlenecks that arise at frequently updated data items (or hot spots ) For some applications, however, the requirement that noncommuting operations must conflict can hurt performance by restricting concurrency. Recently, ....

W.E. Weihl, Commutativity-based concurrency control for abstract data types, 1EEE Trans. Cornput. 37(12) (1988) 1488 1505.

....control in transaction systems today. In recent years much research has focused on extending concurrency control methods to take the semantics of the data into account, thus permitting more concurrency by allowing transactions executing commuting operations to run concurrently (e.g. see [9, 14, 17, 16, 15]) Such logical locking can be important to avoid concurrency bottlenecks that arise at frequently updated data items (or hot spots ) For some applications, however, the requirement that Supported in part by the Defense Advanced Research Projects Agency (DARPA) under Contract ....

W. E. Weihl. Commutativity-based concurrency control for abstract data types. IEEE Transactions on Computers, 37(12):1488-1505, December 1988.

No context found.

. Weihl, W.E.: Commutativity-based concurrency control for abstract data types. IEEE Transactions on Computers, Vol.37, No.12, pages 1488-1505, December 1988

No context found.

W. Weihl. Commutativity-based concurrency control for abstract data types. IEEE Transactions on Computers, 37(12):1488--1505, 1988. 24

No context found.

Weihl, W. E. Commutativity-based concurrency control for abstract data types. IEEE Transactions on Computers 37, 12 (December 1988), 1488-1505.

No context found.

W. Weihl. Commutativity-based concurrency control for abstract data types. IEEE Transactions on Computers, 37(12):1488--1505, December 1988.

No context found.

Weihl, W. E. Commutativity-based concurrency control for abstract data types. IEEE Transactions on Computers 37, 12 (1988), 1488--1505.

No context found.

W. Weihl. Commutativity-based concurrency control for abstract data types. IEEE Transactions on Computers, 37#12#:1488#1505, December 1988.

No context found.

WEIH88 Weihl W. E., Commutativity Based Concurrency Control for Abstract Data Types, IEEE Transactions on Computers 37/12, Dec 1988.

First 50 documents  Next 50

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC